In a previous post we have discussed in some detail the design of a machine learning platform covering model training, serving, governance and observability.
[Read More]
The open-approach to Operationalize Machine Learning
Constructing an open-source ecosystem to govern the complexity of Machine Learning services
This blog post discusses the operationalization of machine learning systems. From an engineering perspective, these systems pose enormous challenges due to the high variability of requirements that different use cases have. As the machine-learning maturity of organizations evolves, the complexity required to manage these systems grows exponentially; The proliferation of...
[Read More]
Proof of concept for constructing ML pipelines
A proposal of a design pattern to support the creation of machine learning pipelines using dataflow and currying.
This post serves as first discussion regarding a proof of concept for what I believe is a proper way to design and manage machine learning pipelines.
[Read More]
Custom Spark ML with a Python wrapper
Extend Spark ML functionality with your own models and provide access from python
Spark is a framework which tries to provides answers to many problems at once. At its core it allows for the distribution of generic workloads to a cluster. But then it provides a SQL-friendly API to work with structured data, a streaming engine to support applications with fast-data requirements and...
[Read More]
Web scrapping with python
A beginners's guide to parse web data
Data is a valuable thing to have. Most of the times it is already out there, laying around in HTML pages or waiting to be requested by a js callback, in what may we can refer to as static and dynamic content, respectively.
[Read More]